- Python < 3.6
- rasterstats
- rasterio
- shapely
- pandas
- numpy
- scipy
- matplotlib
- tqdm
This example shows the phenology of wetland vegetation in the San Francisco delta, California, USA, using three consecutive years of Sentinel-2 data. With a total of ~120 observations per pixel, a tipical temporal figure looks like:
As you can see, if you take only yearly data you may end up with many gaps in the data. So, if you are interested in stable land surface phenology (LSP) metrics, you may want to mix several years together to fill the gaps with other years information:
import phenopy as phen
# coordinates of the pixel to check
X = 584543.89281
Y = 4228482.52520
# plot PhenoShape with multi-year data
phen.PhenoPlot(X=X, Y=Y, inData=inData, dates=dates, ylim=None, type=1,
rollWindow=3, ylab='EVI')
Using multi-year data, PhenoPy is able to fill the gaps and fit a single phenological shape from which estimate phenological parameters, such as the start of the season (SOS), the peak of season (POS), and the end of season (EOS). You can check if these values are correctly assessed by using type=2 in PhenoPlot:
# plot PhenoShape with SOS, POS, and EOS locations
phen.PhenoPlot(X=X, Y=Y, inData=inData, dates=dates, ylim=None, type=1,
rollWindow=3, ylab='EVI')
This is only for visualization. A raster of fitted lines per pixel can be achieve:
phen.PhenoShape(inData=inData, outData=outData, dates=dates,
nan_replace=-32767, nGS=46, chuckSize=256, n_jobs=4)where nan_replace is the NaN value (default None), nGS is the munber of interpolation values to use (here we obtained one avarage value per week, hence = 46), chuckSize is the size of the chunck that rasterio will open at the time, and n_jobs is the number of parallel process to use. We used rasterio to open big images in small-sized chuncks and therefore keep the memory usage low.
Meanwhile, a raster of Land Surface Phenology (LSP) metrics can be obtained as:
phen.PhenoLSP(inData=inData, outData=outData, nGS=46,
n_jobs=4, chuckSize=256)Until now, the metrics included are:
- SOS - DOY of start of season
- POS - DOY of peak of season
- EOS - DOY of end of season
- vSOS - Value at start of season
- vPOS - Value at peak of season
- vEOS - Value at end of season
- LOS - Length of season (DOY)
- AOS - Amplitude of season (in value units)
- IOS - Integral of season (SOS-EOS)
- ROG - Rate of greening [slope SOS-POS]
- ROS - Rate of senescence [slope POS-EOS]
- SW - Skewness of greening period